YOLOv3 custom training and inference using C++
Over the lunar new year I could work on migrating the code from Python to C++. The compilier is simply the Visual Studio VSCC. The open source libraries used are "opencv2" and "libtorch"
Features:
- Custom training
- Inference
- Loading the pretrained YOLOv3 weights
TODO: Training on GPU, distributed training. How to do it with libtorch? (under review)